Abstract:

In time-sensitive application scenarios such as military operations, cloud computing cannot meet users??real-time requirements, so dispersed computing comes into being. It makes use of global computing resources including smart phones, tablets, internet-connected cars and Internet of things terminals to provide services, and regards cloud data centers as common computing nodes, completely eliminating centralization and realizing the decentralization of computing resources. Dispersed computing connects all devices with computing power to form a networked organism, with each computing node providing services to users in a collaborative and shared manner. Different from the localization processing of fog computing and edge computing, this mode makes use of the free computing power in the network, bypasses the limitation of local computing power, and has been widely concerned. This paper first introduces the background of dispersed computing and provides the definition of it. Then this paper introduces three core techniques of dispersed computing in detail. After that, this paper instantiates the concept of dispersed computing through some specific application scenarios, so as to better analyze the advantages of dispersed computing in the Internet of everything era. Finally, this paper presents several challenges and opportunities of dispersed computing.

Key words: dispersed computing, could computing, fog computing, edge computing, Internet of everything

摘要：

在军事作战等对时延敏感的应用场景中，云计算无法满足用户的实时需求，因此分散计算应运而生。它利用智能手机、平板电脑、联网汽车和物联网终端等全球计算资源提供服务，并将云数据中心视为通用计算节点，彻底消除中心化，实现计算资源的分散化。分散计算将所有具有计算能力的设备连接起来，形成一个网络化的有机体，每个计算节点以协作和共享的方式为用户提供服务。与雾计算和边缘计算的本地化处理不同，该范式利用了网络中的空闲计算资源，绕过了局部计算能力的限制，得到了广泛的关注。首先，介绍了分散计算的研究背景，并给出了分散计算的定义；其次，详细介绍了分散计算的三种核心技术；随后，通过一些具体的应用场景实例化分散计算的概念，更好地分析了分散计算在万物互联时代的优势；最后，阐述了未来分散计算的研究方向以及面临的挑战。

关键词: 分散计算, 云计算, 雾计算, 边缘计算, 万物互联